Several publications are referenced in this application by Arabic numerals within parentheses. Full citations for these references are found at the end of the specification immediately preceding the claims. The disclosures of these publications are incorporated by reference into this application in order to more fully describe the state of the art to which this invention pertains.
There is a continuous and expanding need for rapid, highly specific methods of detecting and quantifying chemical, biochemical, and biological substances. Of particular value are methods for measuring small quantities of pharmaceuticals, metabolites, microorganisms and other materials of diagnostic value. Examples of such materials include narcotics and poisons, drugs administered for therapeutic purposes, hormones, pathogenic microorganisms and viruses, antibodies, metabolites, enzymes and nucleic acids.
The presence of these materials can often be determined by binding methods which exploit the high degree of specificity which characterizes many biochemical and biological systems. Frequently used methods are based on, for example, antigen-antibody systems, nucleic acid hybridization techniques, and protein-ligand systems. In these methods, the existence of a complex of diagnostic value is typically indicated by the presence or absence of an observable "label" which has been attached to one or more of the complexing materials. The specific labeling method chosen often dictates the usefulness and versatility of a particular system for detecting a material of interest. A preferred label should be inexpensive, safe, and capable of being attached efficiently to a wide variety of chemical, biochemical, and biological materials without changing the important binding characteristics of those materials. The label should give a highly characteristic signal, and should be rarely, and preferably never, found in nature. The label should be stable and detectable in aqueous systems over periods of time ranging up to months. Detection of the label should be rapid, sensitive, and reproducible without the need for expensive, specialized facilities or personnel. Quantification of the label should be relatively independent of variables such as temperature and the composition of the mixture to be assayed. Most advantageous are labels which can be used in homogeneous systems, i.e., systems in which separation of the complexed and uncomplexed labeled material is not necessary. This is possible if the detectability of the label is modulated when the labeled material is incorporated into a specific complex.
A wide variety of labels have been developed, each with particular advantages and disadvantages. For example, radioactive labels are quite versatile, and can be detected at very low concentrations. However, they are expensive, hazardous, and their use requires sophisticated equipment and trained personnel. Furthermore, the sensitivity of radioactive labels is limited by the fact that the detectable event can, in its essential nature, occur only once per radioactive atom in the labeled material. Moreover, radioactive labels cannot be used in homogeneous methods. Thus, there is wide interest in nonradioactive labels. These include molecules observable by spectrophotometric, spin resonance, and luminescence techniques, as well as enzymes which produce such molecules. Among the useful nonradioactive labeling materials are organometallic compounds. Because of the rarity of some metals in biological systems, methods which specifically assay the metal component of the organometallic compounds can be successfully exploited. For example, Cais, U.S. Pat. No. 4,205,952 (1980) discloses the use of immunochemically active materials labels with certain organometallic compounds for use in quantitating specific antigens. Any general method of detecting the chosen metals can be used with these labels, including emission, absorption and fluorescence spectroscopy, atomic absorption, and neutron activation. The methods often suffer from lack of sensitivity, can seldom be adapted to a homogeneous system, and as with atomic absorption, sometimes entail destruction of the sample.
Of particular interest are labels which can be made to luminesce through photochemical, chemical, and electrochemical means. "Photoluminescence" is the process whereby a material is induced to luminesce when it absorbs electromagnetic radiation. Fluorescence and phosphorescence are types of photoluminescence. "Chemiluminescent" processes entail the creation of the luminescent species by a chemical transfer of energy. "Electrochemiluminescence" entails the creation of the luminescent species electrochemically.
These luminescent systems are of increasing importance. For example, Mandle, U.S. Pat. No. 4,372,745 (1983), discloses the use of chemiluminescent labels in immunochemical applications. In the disclosed systems, the labels are excited into a luminescent state by chemical means such as by reaction of the label with H.sub.2 O.sub.2 and an oxalate. In these systems, H.sub.2 O.sub.2 oxidatively converts the oxalate into a high energy derivative, which then excites the label. This system will, in principle, work with any luminescent material that is stable in the oxidizing conditions of the assay and can be excited by the high energy oxalate derivative. Unfortunately, this very versatility is the source of a major limitation of the technique: typical biological fluids containing the analyte of interest also contain a large number of potentially luminescent substances that can cause high backround levels of luminescence.
Another example of the immunochemical use of chemiluminescence which suffers from the same disadvantages in Oberhardt et al., U.S. Pat. No. 4,280,815 (1981), who disclose the in situ electrochemical generation of an oxidant (e.g., H.sub.2 O.sub.2) in close proximity to an immunoreactant labeled with a chemiluminescent species. The electrogenerated oxidant diffuses to the chemiluminescent species and chemically oxides it, resulting in the net transfer of one or more electrons to the electrogenerated oxidant. Upon oxidation, the chemiluminescent species emits a photon. In contrast, the subject invention requires the direct transfer of electrons from a source of electrochemical energy to a chemiluminescent species which is capable of repeatedly emitting photons.
The present invention is concerned with electrochemiluminescent labels. Suitable labels comprise electrochemiluminescent compounds, including organic compounds and organometallic compounds. Electrochemiluminescent methods of determining the presence of labeled material are preferred over other methods for many reasons. They are highly diagnostic of the presence of a particular label, sensitive, nonhazardous, inexpensive, and can be used in a wide variety of applications. Organic compounds which are suitable electrochemical labels include, for example, rubrene and 9,10-diphenyl anthracene. Many organometallic compounds are suitable electrochemical labels, but of particular use are Ru-containing and Os-containing compounds.
Ru-containing and Os-containing organometallic compounds have been discussed in the literature. Cais discloses that any metal element or combination of metal elements, including noble metals from group VIII such as Ru, would be suitable components of organometallic labels detectable by atomic absorption methods. (Cais, col. 11, line 20). However, ruthenium is not a preferred metal in Cais, osmium is not specifically mentioned, no data are presented on the efficiency of using Ru or Os in any of the methods disclosed, and the preferred method of detection, atomic absorption, entails destruction of the sample.
Weber, U.S. Pat. No. 4,293,310 (1981), discloses the use of Ru-containing and Os-containing complexes as electrochemical labels for analytes in immunoassays. The disclosed complexes are linked to amino groups on the analytes through a linkage. Weber also suggests the possibility of forming carboxylate esters between the labels and hydroxy groups on other analytes.
The presence of the labeled materials can be determined according to Weber, with an apparatus and method which comprises a quencher and an electrochemical flow cell with light means. The photoelectrochemically active label upon photoexcitation transfers an electron to a quencher molecule; the oxidized molecule is subsequently reduced with an electron from an electrode of the flow cell which is held at suitable potential. This electron is measured as photocurrent. The amount of free labeled analyte in the system is determined by the photocurrent signal. This method is the reverse of electrochemiluminescent detection of luminescent materials.
In subsequent reports, Weber et al. discuss the problems associated with the use of this method to detect Ru-containing labels (1). In Table 2 of Weber et al. (1), the extrapolated detection limit for tris(bipyridyl)ruthenium(II) is 1.1.times.10.sup.-10 moles/L under optimal conditions. In anticipating that the actual use of these labels would entail measurements in the presence of complex mixtures, Weber et al. tested for potential interferents in their system. Table 3 of Weber et al. lists dimethylalkyl amines, EDTA, N-methylmorpholine, N,N'-dimethylpiperazine, hydroxide, oxalate, ascorbate, uric acid, and serum as interferents which would presumably raise the practical detection limit substantially above 1.1.times.10.sup.-10 moles/L. These studies were performed with a simple Ru-containing compound. No studies were reported in Weber or Weber et al. regarding the limits of detection of complex substances labelled with Ru-containing labels, or on whether the thiourea linkage between the labeled material and label is stable under conditions of the assay.
The particular labels with which the present invention is concerned are electrochemiluminescent. They can often be excited to a luminescent state without their oxidation or reduction by exposing the compounds to electromagnetic radiation or to a chemical energy source such as that created by typical oxalate-H.sub.2 O.sub.2 systems. In addition, luminescence of these compounds can be induced by electrochemical methods which do entail their oxidation and reduction.
Extensive work has been reported on methods for detecting Ru(2,2'-bipyridine).sub.3.sup.2+ using photoluminescent, chemiluminescent and electrochemiluminescent means (2,3). This work demonstrates that bright orange chemiluminescence can be based on the aqueous reaction of chemically generated or electrogenerated Ru(bpy).sub.3.sup.3+ (where "bpy" represents a bipyridyl ligand) with strong reductants produced as intermediates in the oxidation of oxalate ions or other organic acids. Luminescence also can be achieved in organic solvent-H.sub.2 solution by the reaction of electrogenerated, or chemically generated Ru(bpy).sub.3.sup.1+ with strong oxidants generated during reduction of peroxydisulfate. A third mechanism for production of electrochemiluminescence from Ru(bpy).sub.3.sup.2+ involves the oscillation of an electrode potential between a potential sufficiently negative to produce Ru(bpy).sub.3.sup.1+ and sufficiently positive to produce Ru(bpy).sub.3.sup.3+. These three methods are called, respectively, "oxidative reduction," "reductive oxidation," and "the Ru(bpy).sub.3.sup.3+/+ regenerative system."
The oxidative-reduction method can be performed in water, and produces an intense, efficient, stable luminescence, which is relatively insensitive to the presence of oxygen or impurities. This luminescence from Ru(bpy).sub.3.sup.2+ depends upon the presence of oxalate or other organic acids such as pyruvate, lactate, malonate, tartrate and citrate, and means of oxidatively producing Ru(bpy).sub.3.sup.3+ species. This oxidation can be performed chemically by such strong oxidants as PbO.sub.2 or a Ce(IV) salt. It can be performed electrochemically by a sufficiently positive potential applied either continuously or intermittently. Suitable electrodes for the electrochemical oxidation of Ru(bpy).sub.3.sup.2+ are, for example, Pt, pyrolytic graphite, and glassy carbon. Although the oxalate or other organic acid is consumed during chemiluminescence, a strong, constant chemiluminescence for many hours can be achieved by the presence of an excess of the consumed material in the reaction chamber.
The reductive-oxidation method can be performed in partially aqueous solutions containing an organic co-solvent such as, for example, acetonitrile. This luminescence depends upon the presence of peroxydisulfate and a means of reductively producing Ru(bpy).sub.3.sup.1+ species. The reduction can be performed chemically by strong reductants such as, for example, magnesium or other metals. It can be performed electrochemically by a sufficiently negative potential applied either continuously or intermittently. A suitable electrode for the electrochemical reduction of Ru(bpy).sub.3.sup.2+ is, for example, a polished glassy-carbon electrode. As with the oxidative-reduction method, continuous, intense luminescence can be achieved for many hours by inclusion of excess reagents, or by continuous addition of the consumed reagents to the reaction mixture.
The Ru(bpy).sub.3.sup.3+/+ regenerative system can be performed in organic solvents such as acetonitrile or in partially aqueous systems, by pulsing an electrode potential between a potential sufficiently negative to reduce Ru(bpy).sub.3.sup.2+ and a potential sufficiently positive to oxidize Ru(bpy).sub.3.sup.2+. A suitable electrode for such a regenerative system is, for example, a Pt electrode. This system does not consume chemical reagents and can proceed, in principle, for an unlimited duration.
These three methods of producing luminescent Ru-containing compounds have in common the repetitive oxidation-reduction or reduction-oxidation of the Ru-containing compound. The luminescence of solutions containing these compounds is therefore highly dependent on the electric potential of the applied energy source, and is therefore highly diagnostic of the presence of the Ru-containing compound.
Mandle cites Curtis et al. (4) as a possible label in chemiluminescent applications. Curtis et al. reports only unpublished observations that Ru complexes can be induced to emit light when chemically excited by an oxalate/H.sub.2 O.sub.2 system (Curtis et al., p. 350). Neither Mandle nor Curtis recognized the exceptional utility of ruthenium and osmium complexes in chemiluminescent applications or the utility of electrochemiluminescent systems.
Sprintschnik, G., et al. (5), have described complexes of tris (2,2'-bipyridine)ruthenium(II) esterified with octadecanol or dehydrocholesterol, and have created monolayer films of these surfactant complexes. The complexes were photoluminescent. But when the films were exposed to water, and then to light, the Ru-complexes failed to photoluminesce. This was attributed to photohydrolysis of ester groups in the presence of light.
As described in parent applications Ser. Nos. 858,354 and PCT 87/00987, a variety of analytes of interest and chemical substances that bind to analytes of interest may be conveniently attached to Ru-containing or Os-containing labels through amide or amine linkages. The labeled materials may then be detected by any of a wide variety of means, but by far the most efficient, reliable, and sensitive means are photoluminescent, chemiluminescent, and electrochemiluminescent means. Electrochemiluminescent labels, including Ru-containing and Os-containing labels are particularly versatile and advantageous.
For many years the food industry has been concerned with the presence of biological and chemical contaminants in raw food components and processed foods. While technological advances have been made in reducing the occurrence of food contamination and food borne disease outbreaks resulting therefrom, little progress has been reported in developing rapid and sensitive methods for the detection and identification of food contaminants. Existing standard methods for the detection of harmful contaminants in foods are generally very time consuming, labor intensive, and technically difficult. While the analytical methods themselves for the most part are of adequate sensitivity, the lengthy sample preparation procedures prior to the performance of the detection method often results in low yield of the contaminant in question so that false negatives are frequently encountered.
Two examples which serve to illustrate these problems are the currently recognized standard methods for detecting the presence of Salmonella and Staphylococcal enterotoxins in foods. The detection of Salmonella in food involves several enrichment stages due to the fact that these bacteria, when present in foods, are usually found in low numbers and are often sublethally injured. Therefore, detection methods for Salmonella must be sensitive and allow for the resuscitation and growth of injured cells.
Two methods for Salmonella detection are currently recommended by the U.S. Food and Drug Administration. The methods appear in The Bacteriological Analytical Manual for Foods, 6th ed., Association of Official Analytical Chemists, Washington, D.C. (1984). One method is a pure culture technique involving preenrichment, selective enrichment and selective plating, a procedure which requires 4 days to obtain presumptive results and 5 to 7 days to obtain complete results. The other method involves immunofluorescence after selective enrichment. This procedure is more rapid, however it can results in a high incidence of false positive results due to problems of cross reactivity of the polyvalent antisera used in the test (6, 7).
The procedure recommended by the U.S. Food and Drug Administration for the detection of Staphylococcal enterotoxins in foods also appears in The Bacteriological Analytical Manual for Foods, 6th ed., Association of Official Analytical Chemists, Washington, D.C. (1984). This method involves the concentration of an extract of a large food sample, e.g., approximately 100 grams, to a small volume, e.g., approximately 0.2 ml, by several dialysis concentration steps and an ion exchange column purification of the sample extract in order to prepare the sample for the microslide double-immunodiffusion technique. This procedure generally requires more than a week to perform.
Tests which are more rapid have recently been developed for the detection of a variety of contaminants such as bacteria, toxins, antibiotics and pesticide residues in foods. In many cases however, sample preparation prior to running the assay continues to be laborious and time consuming. Radioimmunoassays (RIA) and enzyme-linked immunosorbent assays (ELISA) have shortened the hands-on time for the analytical method itself, however these methods are still labor intensive and far from simple to perform. In addition, these methods are usually based on the use of polyclonal antisera, which are variable to specificity and sensitivity, and are generally in short supply for testing for a given food contaminant. ELISA methods have been developed for the analysis of food samples which employ monoclonal antibodies rather than polyclonal antisera. The use of monoclonal antibodies in an assay system for a food contaminant assures the constant supply of reagent which imparts unchanging specificity and sensitivity to the test itself. Monoclonal antibodies have been used in ELISA systems to test for food contaminants such as Salmonella (8) and Staphylococcal enterotoxins (9). Commercially available products for Salmonella detection which employ EIA methodology (Bio-Enzabead Screen Kit, Litton Bionetics) and DNA probe technology (Gene-Trak, Integrated Genetics) are time consuming and labor intensive. Commercially available tests for detection of Staphylococcal enterotoxin in foods which employ reversed passive latex agglutination (SET-EIA, Dr. W. Seiken Co.) and EIA methodology (SET-RPLA, Denka Seiken Co.) and EIA methodology (SER-EIA, Dr. W. Bommeli Laboratories) suffer from the same limitations.
For the past 100 years the bacterium Escherichia coli and the coliform group have been commonly used as indicators to monitor water quality and incidences of sewage contamination.
Current detection methodologies for E. coli and/or coliforms are based on the properties of acid or gas production from the fermentation of lactose. The most widely used methods are: the Most Probable Number (MPN) assay and the Membrane Filtration (MF) test. Both techniques are approved by the Environmental Protection Agency (EPA) and the American Public Health Association (APHA) for the microbiological examination of water and waste water (10), and also by the Food and Drug Administration (FDA) for the bacteriological examination of milk and foods (11).
The MPN method is actually comprises of three (12) separate assays (10). In the Presumptive test, a nonselective medium such as Lauryl Sulfate Tryptose (LST) broth or Lactose broth is used to check for gas production from the fermentation of lactose. Gas positive tubes are then subcultured into a more selective medium, Brillant Green Lactose Bile (BGLB) broth for coliforms and E. coli (EC) broth for fecal coliforms, and again checked for gas production (confirmed test). Samples from positive Confirmatory tests are required to be tested further by plating on a selective and differential medium like Eosin Methylene Blue (EMB) agar or Endos agar, followed by Gram Stain and some biochemical tests to firmly establish the presence of the indicator bacteria (Completed test). The entire MPN assay may require up to five (5) days for completion; therefore, for routine water analysis, most laboratories use only the Presumptive and the Confirmed portions of the MPN assay, which still requires 48 hours to 72 hours to complete. In addition to being time consuming and cost ineffective in terms of materials, incidences of both false positive and false negative reactions have also been commonly encountered in the MPN assays (15, 16, 20).
The MF technique for the bacteriological examination of water was introduced in the early 1950's (12). Unlike the MPN assay, which was tedious and time consuming, MF analysis could be complete in 24 hours without the need for further confirmations. The basic MF procedure is as follows: A volume of water sample, usually 100 ml is filtered through a 0.45 um pore diameter filter, and then incubated on a sterile pad saturated with selective medium. The two media most often used are the mEndo broth, selective for coliforms at 35.degree. C., and the mFC broth, selective for fecal coliforms at 44.5.degree. C. (10). Since the introduction of the media, numerous authors have reported that both the mEndo and the mFC broth tends to underestimate the actual numbers of indicator bacterial present, due either to the selectivity of the medium or the high temperature used for incubation (44.5.degree. C.) (21, 22). Such incidences of false negatives have been especially prevalent when the organisms in the sample have been sublethally injured by environmental factors and/or chemicals (17, 18). Recently, modifications have been proposed by the EPA to follow up the MF test by a confirmatory procedure, whereby at least ten colonies on each filter need to be checked for gas production using the LST broth followed by BGLB broth as in the MPN assay (14). Such modifications although would reduce the incidences of both false negative and false positive reactions, it would also increase material cost as well as triple the MF assay time from 24 hours to 72 hours.
In 1982, Feng and Hartman introduced a fluorogenic assay for the detection of E. coli using the substrate 4-methyl umbelliferone glucuronide (MUG) (13). E. coli cells produced the enzyme beta-glucuronidase which would cleave the substrate releasing the fluorogenic 4-methylumbelliferone radical (19). By incorporating the compound MUG into the Presumptive LST medium, a single tube of LST-MUG medium provided both the Presumptive data (gas production) and the Confirmed data (fluorescence) for fecal coliforms within 24 hours. Although the MUG assay was rapid and simple, only 85% to 95% of the E. coli (depending on source) produced this enzyme, hence the test was not 100% reliable. Also the system was not applicable to the coliform group.
Currently, no suitable assay exists for the detection and enumeration of coliforms and fecal coliforms in a sample. The development of a simple, rapid, and reliable detection assay would not only decrease cost and time, but also greatly increase the efficiency of monitoring water sanitation and food processing and handling.